Predicting Bioconcentration Factor Using a Metabolism-Based Quantitative Structure-Activity Relationship Model Open Access
Downloadable ContentDownload PDF
The prediction of bioconcentration factor is important for the determination of potential environmental toxicity of new and existing chemical substances. To this end, various in silico models for estimating bioconcentration factor have been developed to replace or assist in vivo testing, including linear and nonlinear Quantitative Structure-Activity Relationship Models (QSARs). This work is exceptional among these models because it is the first to consider metabolic effects relating to Cytochrome P450 in fish a priori to model development, as opposed to establishing metabolic explanations for model behavior a posteriori. The model developed here is based on a subset of data from previous work by Dimitrov et al (Dimitrov, Dimitrova et al. 2005). The final model presented will be a two-step process. First, structures are classified into a potential metabolism category using SMARTS patterns encompassing known functional domains for reactivity or metabolism by Cytochrome P450. Then a linear QSAR model utilizing calculated chemical descriptors specific to the established potential metabolism category is applied to the compound to estimate bioconcentration factors. It is expected that this metabolism-based QSAR model will find applications in regulatory settings and will be used as a basis for expansion to other chemical classes outside of those considered here.